Geophysical methods for tunnel detection and characterization of a coal mine in Ermelo

• Main Author: Moleleki, Nkimo Stephen
• Format: Others
• Language: en
• Published: 2020
• Online Access: https://hdl.handle.net/10539/29620

A dissertation submitted to the Faculty of Science, University of the Witwatersrand, in fulfilment of the requirements for the degree of Master of Science, Johannesburg November 2019 === Void detection with gravity, DC resistivity and Ground Penetrating Radar (GPR) have been ubiquitously documented. Two techniques, namely magnetic and time-domain electromagnetic (TDEM), have received less attention. This research was aimed at evaluating the capability of four geophysical methods, namely gravity (specifically microgravity) as benchmark well known for tunnel and subsurface cavity detection, magnetics and time-domain electromagnetic (TDEM) methods, as well as a low frequency broad band ground penetrating radar (GPR) to delineate shallow mine adits (tunnels), characterize their physical properties (dimensions, strike, whether air-filled or water-filled) and depth. However, an old colliery on the farm called Driehoek in the Ermelo district discussed in this document was only accessible for a very limited time period as open cast mining was to commence and the low frequency GPR had to be abandoned but was used to delineate mine workings at a depth of 40 m at another site (Vlakhoogte Coal Mine near Delmas in Mpumalanga) and is briefly reported on. Thus, the main geophysical methods being evaluated were reduced to three. A suitable study area available with shallow tunnel development was this old colliery underlain mostly by electrically resistive sandstones of ~180 ohm.m, with more conductive minor shales and siltstones as country rock and coal seams exploited by checkerboard and bord and pillar mining method at shallow depth of ~5 m from surface. Both magnetic and microgravity data were collected on the same grid pattern overlaying each other, but the TDEM consisted of a single line cross-cutting the microgravity grid pattern at an angle of about 35 degrees from the north. Evidence of underground openings was seen from a reconnaissance resistivity survey and a TDEM apparent resistivity pseudo section (resistivity values from the TDEM were orders of magnitude higher (~750 ohm.m) than the surrounding rock (~180 ohm.m)). 2D modelling of the microgravity and magnetic datasets was undertaken by using the free-ware Grav2dc and Mag2dc programs respectively. 3D density modelling however revealed that the openings are not tunnels but rather openings that will be referred to as bords/cavities. The results of 2D modelling of microgravity and magnetic datasets revealed that the bords are buried ~5 m from surface, have a width of ~3 m and have a stope width of 2 m. Forward modelling software, GmodIntrp.for and MagModIntrp.for, were used to produce 3D density and susceptibility models of the subsurface from the microgravity and magnetic data respectively and displayed in Geosoft Oasis Montaj. From the 3D density voxel model, a density contrast of these voids/bords of -2.47 g/cm3 with regards to the surrounding sandstone gave modelling results close to the theoretical value of -2.50 g/cm3 for air-filled cavities leading to the conclusion that the bords do not contain water. The 3D models did not reveal the strike of these bords although some were inferred to be interconnected from the visualization of the 3D density model. The success of the study could be bolstered beyond doubt through producing a conclusive 3D geological model that is a representation of the subsurface incorporating 3D models of all the datasets and not just microgravity and magnetics. The final conclusion is that the three methods under study can be successfully used to delineate openings (bords, drives, galleries, shafts and portals) at coal mines and to contribute towards reducing the cost of drilling by pinpointing where these openings can be intercepted. Knowing the lithology and its respective physical properties in the area thoroughly will contribute to deciding which geophysical methods are most applicable to address the problem. === XN2020